FORWARD GUIDANCE AND HETEROGENEOUS BELIEFS. Discussion by Yuriy Gorodnichenko (UC Berkeley) Philippe Andrade Gaetano Gaballo Eric Mengus Benoit Mojon

Transcription

1 FORWARD GUIDANCE AND HETEROGENEOUS BELIEFS Philippe Andrade Gaetano Gaballo Eric Mengus Benoit Mojon Discussion by Yuriy Gorodnichenko (UC Berkeley) 1

2 GREAT RECESSION & MONETARY POLICY Conventional policy in recessions: o stimulate the economy by cutting interest rates 2

3 GREAT RECESSION & MONETARY POLICY Conventional policy in recessions: o stimulate the economy by cutting interest rates Zero lower bound constraint non-conventional monetary policy o Forward guidance: Promise to keep interest rates low in the future to generate inflation expectations and stimulate the economy now 3

4 GREAT RECESSION & MONETARY POLICY Conventional policy in recessions: o stimulate the economy by cutting interest rates Zero lower bound constraint non-conventional monetary policy o Forward guidance: Promise to keep interest rates low in the future to generate inflation expectations and stimulate the economy now A key challenge: interpretation of the Fed s announcements o A promise indicates that the current fundamentals are bad and the Fed reacts to the poor economic conditions ( delphic ; state of the economy ) o A promise indicates an innovation to monetary policy ( odyssean, stance of monetary policy ) 4

5 COMMUNICATION: RISING COMPLEXITY Source: Hernández-Murillo and Shell (St. Louis Fed s Economic SYNOPSES, Nov 2014) 5

6 COMMUNICATION: FED WATCHERS Source: WSJ Fed Statement Tracker 6

7 COMMUNICATION: WSJ (FEB 5, 2015) WSJ: We re all a little consumed right now with the word patient in the policy statement, which I m sure really pleases you. What will it mean when the Fed removes patient from its statement? PLOSSER: It could mean different things to different people. The chair has articulated what she means. I think it is fair to say that once the committee chooses to remove patient that they are making a different statement about the likelihood and timing of rate increases. More precisely what that means, it will depend on economic conditions. We would save ourselves a lot of headaches and you guys writing a lot of words if we didn t use (words like patient ). I ve been only marginally successful in reducing our efforts to do that. WSJ: The Fed added new words to its policy statement in January, a reference to international developments. PLOSSER: Our statement is a bit like Hotel California. Words check in and they never check out. We have way too many words. I think our statement is too long. I think it is too confusing. I think it needs a thorough rewrite. I ve suggested that before. Over the course of this unusual period of time, the statement has gotten longer and longer and policy has gotten more and more complex. I m not sure that all of the extra words and all of the complexity have necessarily improved our communication particularly. At times I think it has gotten more confusing and not less confusing. Source: Plosser Says Fed Will Struggle With Policy Statement at March Meeting 7

8 FORWARD GUIDANCE AND HETEROGENEOUS BELIEFS Policy rule: Ω Ω fundamentals policy shock 8

9 FORWARD GUIDANCE AND HETEROGENEOUS BELIEFS Policy rule: Ω Ω fundamentals policy shock Economic agents: Ω 9

10 FORWARD GUIDANCE AND HETEROGENEOUS BELIEFS Policy rule: Ω Ω fundamentals policy shock Economic agents: Ω Agents observe only and have to make inference about Ω and 10

11 FORWARD GUIDANCE AND HETEROGENEOUS BELIEFS Policy rule: Ω Ω fundamentals policy shock Economic agents: Ω Agents observe only and have to make inference about Ω and Two interpretations of news about low future : Ω is bad and the Fed reacts to the poor fundamentals ( delphic ) Consume less now is low which indicates expansionary policy of the Fed ( odyssean ) Inflation expectations Consume more now 11

12 FORWARD GUIDANCE AND HETEROGENEOUS BELIEFS Policy rule: Ω Ω fundamentals policy shock Economic agents: Ω Agents observe only and have to make inference about Ω and Two interpretations of news about low future : Ω is bad and the Fed reacts to the poor fundamentals ( delphic ) Consume less now is low which indicates expansionary policy of the Fed ( odyssean ) Inflation expectations Consume more now How to measure? Examine disagreement Ω 12

13 DISAGREEMENT Inflation t+4,t Interest rate rt+4 fixed date flexible date Survey of Professional Forecasters: Interquartile range

14 FORWARD GUIDANCE AND COVARIANCE OF BELIEFS Policy rule: Ω Ω fundamentals policy shock Economic agents: Ω Ω 14

15 FORWARD GUIDANCE AND COVARIANCE OF BELIEFS Policy rule: Ω Ω fundamentals policy shock Economic agents: Ω Ω Two interpretations of news about low future : Ω is bad and the Fed reacts to the poor fundamentals ( delphic ) If forecaster A has a more negative belief about than forecaster B ( ), then forecaster A should predict a worse fundamental than forecaster B ( Ω Ω ). In the cross-section, the covariance between Ω and should be POSITIVE. 15

16 FORWARD GUIDANCE AND COVARIANCE OF BELIEFS Policy rule: Ω Ω fundamentals policy shock Economic agents: Ω Ω Two interpretations of news about low future : Ω is bad and the Fed reacts to the poor fundamentals ( delphic ) If forecaster A has a more negative belief about than forecaster B ( ), then forecaster A should predict a worse fundamental than forecaster B ( Ω Ω ). In the cross-section, the covariance between Ω and should be POSITIVE. is low which indicates expansionary policy of the Fed ( odyssean ) If forecaster A believes in expansionary policy more strongly than forecaster B ( ), then forecaster A should predict a stronger fundamental than forecaster B ( Ω Ω ). In the cross-section, the covariance between Ω and should be NEGATIVE. 16

17 FORWARD GUIDANCE AND COVARIANCE OF BELIEFS Procedure: Estimate policy reaction function using pre-great Recession period (1992Q1-2007Q1),,,,, Calculate residuals,. Examine time series of,,, and,,,. 17

18 CORRELATION BETWEEN MP SHOCKS AND INFLATION correlation (, MP ) fixed date flexible date If inflation is a good fundamental, then a negative correlation indicates that agents interpret forward guidance as revealing new information about monetary policy ( odyssean ) rather than fundamentals ( delphic ). 18

19 CORRELATION BETWEEN MP SHOCKS AND (MINUS) UNEMPLOYMENT correlation (-UE, MP ) fixed date flexible date Low unemployment is a good fundamental. Hence a negative correlation indicates that agents interpret forward guidance as revealing new information about monetary policy ( odyssean ) rather than fundamentals ( delphic ). 19

20 ADDITIONAL COMMENTS Forward guidance can work via other channels (e.g., reduce uncertainty). 20

21 ADDITIONAL COMMENTS Uncertation about PCE inflation rate fixed date flexible date Forward guidance can work via other channels (e.g., reduce uncertainty). 21

22 ADDITIONAL COMMENTS Forward guidance can work via other channels (e.g., reduce uncertainty). o Some support If one is interested in which view prevails (delphic or odyssean), looking at the first moments may be more informative. 22

23 Probability of output contracting next year (SPF) ADDITIONAL COMMENTS fixed date flexible date First moments in the SPF did not move much at the time announced forward guidance. 23

24 ADDITIONAL COMMENTS Forward guidance can work via other channels (e.g., reduce uncertainty). o Some support If one is interested in which view prevails (delphic or odyssean), looking at the first moments may be more informative. o First moments did not move much at the time announced forward guidance. The model uses the Survey of Professional Forecasters while the model is about firms. 24

25 ADDITIONAL COMMENTS MSC: share SPF: change size MSC: Share of HH expectting interest rate to rise SPF: Ave. expected size of interest rate change; right axis SPF and Michigan Survey of Consumers suggest different timing of lift-off in the interest rates. 25

26 ADDITIONAL COMMENTS Forward guidance can work via other channels (e.g., reduce uncertainty). o Some support If one is interested in which view prevails (delphic or odyssean), looking at the first moments may be more informative. o First moments did not move much at the time announced forward guidance. The model uses the Survey of Professional Forecasters while the model is about firms. o SPF and Michigan Survey of Consumers suggest different timing of lift-off in the interest rates. The difference between delphic and odyssean views only matters if economic agents with stronger inflationary expectations consume more. o Bachmann, Berg, and Sims (2015) find no evidence for this in the Michigan Survey of Consumers. 26

27 SUMMARY Excellent question and great start. Using alternative moments of cross-sectional dispersion in forecasts could be more informative about how agents interpret forward guidance. Using additional moments of SPF and other sources of expectations can help assess the power of forward guidance. 27